An image denoising algorithm based on partial pixel characteristics
Xie Key · Applied science and technology · 2015
Considering the negative effect of the noise of an electric multiple unit( EMU) image taken by the linear array cameras on the image identification results,this paper proposes an image denoising algorithm based on partial pixel characteristics. Firstly,this algorithm will scan the pixels of the image to be processed line by line,then calculate the mean value of pixels of each line,and at the same time,record the mean value of the pixels of the adjacent equidistant lines up and down. Furthermore,the mean value of pixels of the line is compared with the mean value of the equidistant lines down. If the value is larger than the set threshold,then assign the mean value of the pixels of the equidistant lines up and down to every pixel point corresponding to the line; on the contrary,the pixel value of the line should be kept invariant. By comparing with the classic denoising algorithms,this new algorithm is proven to be effective. This algorithm was applied to the pre-treatment of the EMU image identification system,showing that this algorithm could greatly lower the rate of false identification caused by the noise.